1823.] Sulphuric Acid, Peroxide of Iron, and Ammonia. 407 



acid and precipitated by ammonia weighed in the first experi- 

 ment 16*37. From 50 parts of the salt in two subsequent expe- 

 riments, I obtained 8*32 and 8*2, the mean of which is 8-233 = 

 16-47 per cent. Fifty parts of the salt were exposed for three 

 days to a temperature near that of boiling water ; when 

 weighed, they had lost 21-74 = 43-48 per cent.; the remaining 

 powder was readily dissolved in water, excepting a very small 

 quantity of a brownish-red substance. 



I found great difficulty in ascertaining the exact quantity of am- 

 monia contained in this salt. I dissolved 100 parts in water, added 

 caustic potash to the solution, and subjected it to distillation, so 

 as to pass the vapour into a solution of nitrate of lead ; although 

 the subnitrate thus precipitated by ammonia is nearly insoluble 

 in a solution of neutral nitrate of lead ; it is, however, soluble 

 in water, and when I washed the filter on which it was collected, 

 it almost entirely disappeared. I attempted to expose these 

 solutions of the subnitrate to the action of carbonic acid, which 

 decomposes the subnitrate into carbonate and neutral nitrate, 

 and then the quantity of carbonate of lead would be proportional 

 to the quantity of ammonia passed through a neutral solution of 

 the nitrate ; this method did not, however, succeed much 

 better, nor was nitrate of mercury employed with much greater 

 advantage. At length on comparing the numbers for sulphuric 

 acid and oxide of iron, I found that if I considered the persul- 

 phate to be composed, according to the proportions indicated 

 by Berzelius as constituting what he terms sulphas ferricum, one- 

 fourth of the sulphuric acid remained uncombined with oxide, 

 and, I conceive, this quantity to be combined with ammonia. 



One hundred and fifty parts of the salt were dissolved in 

 water in a flask, and mixed with potash sufficient to decom- 

 pose the whole ; a glass tube was adapted to the flask, and 

 immersed into a solution of 50 parts of the salt in question; the 

 mixture in the flask was boiled for half an hour, and eveiy pre- 

 caution was taken to prevent the escape of any ammonia. At 

 the close of the experiment, the solution through which the 

 ammonia had passed was colourless ; it slightly restored the 

 blue colour of reddened litmus paper, but this effect did not take 

 place unless the paper remained for some time in the solution ; 

 the precipitate had a yellowish colour indicating an admixture of 

 subpersulphate of iron with the peroxide ; the solution and pre- 

 cipitate were heated in a close vessel for 24 hours to a temper- 

 ature of about 100°; the yellow colour of the precipitate had 

 now disappeared, and it seemed to be pure peroxide of iron ; the 

 alkali of the solution was also so perfectly expelled that it did 

 not produce any effect upon reddened litmus paper ; but when 

 mixed with a reddened solution of litmus, a very slight trace of 

 alkali was discoverable. 



The oxide of iron was washed until the water passing through 

 the filter ceased to affect the solution of nitrate of barytes ; the 



